[InferDataLayouts] Add fallback for layout propagation

If there is no tensor layout annotation for the input of a non-FINN
operator, annotate with the same fallback defaults as for the FINN-ops
instead of propagating empty/invalid layouts.

src/qonnx/transformation/infer_data_layouts.py

@@ -62,9 +62,23 @@ def _dims_to_layout(model, node, ndims):
                 else:
                     return DataLayout.UNKNOWN
         else:
-            # propagate input layout to output
-            # TODO this won't work for concat, squeeze/unsqueeze/reshape...
-            return model.get_tensor_layout(node.input[0])
+            # Check whether there is a layout annotation for the first input
+            # TODO: There are multi-input operations, why should the first
+            #  determine the output layout?
+            if layout := model.get_tensor_layout(node.input[0]):
+                # If annotation present: propagate input layout to output
+                # TODO: this won't work for concat, squeeze/unsqueeze/reshape...
+                return layout
+            # Fallback to the same defaults as for the FINN-Ops above
+            else:
+                if ndims == 4:
+                    return DataLayout.NHWC
+                elif ndims == 3:
+                    return DataLayout.NWC
+                elif ndims == 2:
+                    return DataLayout.NC
+                else:
+                    return DataLayout.UNKNOWN
 
 
 def _infer_node_data_layout(model, node):